Control method and system of nursing robot

Abstract

The invention relates to the technical field of nursing robot control, in particular to a control method and system of a nursing robot. The method comprises the following steps that a 3D camera is used for collecting a posture track when a patient gets up, muscle tension change signals associated with the posture track are obtained through an electromyographic signal sensor, and posture associated muscle tension data are obtained through synchronous association processing. Thirdly, the data are analyzed, the muscle group force line unbalance strength is evaluated, and the gravity center shift trend is obtained in combination with gravity center shift regression analysis; finally, robot auxiliary output parameters are generated based on the gravity center shift regression trend and input into a nursing robot control framework, a control scheme suitable for the nursing robot is designed, a control instruction is sent to a robot terminal, and corresponding nursing operation is executed. According to the invention, the control technology of the nursing robot is optimized, so that the control technology of the nursing robot is more perfect.

Description

Technical Field

[0001] The present invention relates to the technical field of nursing robot control, and in particular to a control method and system for a nursing robot. Background Art

[0002] Existing nursing robots primarily rely on simple programmed control and basic sensor feedback to operate, but this approach has certain limitations. For example, during processes such as when a patient stands up or changes position, traditional robots are unable to accurately perceive the patient's real-time physical condition and needs. This results in inaccurate robot-controlled movements and an inability to make timely adjustments based on the patient's specific physical condition, potentially causing discomfort or injury to the patient. Therefore, intelligent nursing robots need to possess higher-level control capabilities that can not only perceive changes in the patient's posture, but also analyze physiological signals such as muscle tension and center of gravity shift in real time to more accurately assist the patient's movements. However, a traditional control method for nursing robots is unable to accurately perceive changes in the center of gravity and fluctuations in muscle tension when the patient stands up during the nursing process, resulting in large control errors in the nursing robot during the assistive support process. Summary of the Invention

[0003] Based on this, it is necessary to provide a control method and system for a nursing robot to solve at least one of the above technical problems.

[0004] To achieve the above object, a control method for a nursing robot is provided, the method comprising the following steps: Step S1: using a 3D camera to collect a patient's posture trajectory when the patient stands up during the nursing process; using an electromyographic signal sensor and based on the posture trajectory, collecting a muscle tension change signal of the body contact area when the patient stands up during the nursing process; performing synchronous correlation processing on the muscle tension change signal based on the posture trajectory to obtain posture-correlated muscle tension change data; Step S2: analyzing the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data, and then performing center of gravity offset regression deconstruction to obtain a center of gravity offset regression trend; generating robot-assisted output parameters based on the center of gravity offset regression trend; Step S3: The robot-assisted output parameters are used as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture, and the architecture is sent to the control terminal of the nursing robot to execute the control of the nursing robot.

[0005] Preferably, the present invention further provides a control system of a nursing robot, for executing the control method of the nursing robot as described above, the control system of the nursing robot comprising: The data acquisition module is used to collect the posture trajectory of the patient when standing up during the nursing process through a 3D camera; collect muscle tension change signals of the body contact area when the patient stands up during the nursing process through an electromyographic signal sensor and based on the posture trajectory; and synchronously correlate the muscle tension change signals based on the posture trajectory to obtain posture-related muscle tension change data; The robot output parameter analysis module is used to analyze the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data. Subsequently, it performs center of gravity offset regression deconstruction to obtain the center of gravity offset regression trend; and generates robot-assisted output parameters based on the center of gravity offset regression trend. The control architecture design module is used to take the robot-assisted output parameters as input to design the control architecture of the nursing robot, so as to build a robot-assisted nursing architecture, and send it to the control terminal of the nursing robot to execute the control of the nursing robot.

[0006] The present invention provides a beneficial effect by combining a 3D camera with an electromyographic (EMG) sensor to capture, in real time, the patient's posture trajectory and muscle tension change signals as they rise during nursing care. This process ensures comprehensive awareness of the patient's physical condition, particularly during dynamic changes, accurately capturing subtle variations in each patient's movements. Synchronously correlating and processing this data generates highly accurate posture-related muscle tension change data, providing a reliable foundation for subsequent data analysis and control decisions. This high-precision data collection and processing enables the robot to precisely respond to the patient's actual needs, enhancing the intelligence and personalization of nursing services. In-depth analysis of posture-related muscle tension change data assesses the strength of muscle group force imbalances during posture changes. This analysis reveals the mechanical state of various body parts during the patient's rise, enabling timely identification of postural imbalances or maladaptation. Further center of gravity offset regression decomposition allows the solution to determine the patient's center of gravity change trend during the rise, accurately predicting the patient's stability and risk during movement. This process provides critical physiological data support, enabling the robot to dynamically adjust to the patient's posture and center of gravity changes, ensuring safety and comfort during movements such as rising. Based on the robot-assisted output parameters generated in the first two steps, an adaptive nursing robot control architecture was designed. This architecture enables dynamic control based on the patient's real-time physiological state (such as posture changes, muscle tension, and center of gravity shift). By receiving these output parameters, the robot adjusts its movements to suit the patient's individual needs, accurately assisting the patient in movements such as rising and adjusting their position. This control architecture enables the robot to provide precise assistance in complex nursing scenarios, avoiding discomfort or injury to the patient caused by mechanical movements. Through this adaptive control, the nursing robot not only improves the safety of nursing services, but also enhances the efficiency and quality of care, enabling its application in a wider range of clinical nursing scenarios. Therefore, this invention optimizes a conventional nursing robot control method, resolving the problem that conventional nursing robot control methods cannot accurately perceive changes in the center of gravity and muscle tension fluctuations when the patient rises during nursing care, resulting in large control errors during the nursing robot's assistance process. This improves the accuracy of sensing changes in the center of gravity and muscle tension fluctuations when the patient rises during nursing care, reducing control errors during the nursing robot's assistance process. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 A schematic flow chart of a control method for a nursing robot; Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG. Figure 3 for Figure 2 Detailed implementation steps of step S22 are shown in the flowchart. DETAILED DESCRIPTION

[0008] See also Figures 1 to 3 , a control method for a nursing robot, the method comprising the following steps: Step S1: using a 3D camera to collect a patient's posture trajectory when the patient stands up during the nursing process; using an electromyographic signal sensor and based on the posture trajectory, collecting a muscle tension change signal of the body contact area when the patient stands up during the nursing process; performing synchronous correlation processing on the muscle tension change signal based on the posture trajectory to obtain posture-correlated muscle tension change data; Step S2: analyzing the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data, and then performing center of gravity offset regression deconstruction to obtain a center of gravity offset regression trend; generating robot-assisted output parameters based on the center of gravity offset regression trend; Step S3: The robot-assisted output parameters are used as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture, and the architecture is sent to the control terminal of the nursing robot to execute the control of the nursing robot.

[0009] In the embodiment of the present invention, reference Figure 1 The above is a flow chart of the steps of a control method of a nursing robot according to the present invention. In this example, the control method of the nursing robot includes the following steps: Step S1: using a 3D camera to collect a patient's posture trajectory when the patient stands up during the nursing process; using an electromyographic signal sensor and based on the posture trajectory, collecting a muscle tension change signal of the body contact area when the patient stands up during the nursing process; performing synchronous correlation processing on the muscle tension change signal based on the posture trajectory to obtain posture-correlated muscle tension change data; In this embodiment of the present invention, a 3D camera is mounted on the head of the nursing robot. Based on the principle of structured light imaging, this camera continuously captures spatial point cloud data at a frame rate of 30Hz. Each frame contains 200,000 3D points, and the measurement range is set within a spatial area of ​​1.2m × 0.8m × 1.5m. This range covers the entire motion trajectory of a patient from sitting to standing on the edge of a nursing bed. Simultaneously, an integrated electromyographic (EMG) sensor array is attached to the inner contact surface of the palm of the nursing robot's robotic arm. Each sensor array consists of four electrodes, with a spacing of 20mm between electrodes. Each electrode point corresponds to a muscle group at the patient's contact site, including the forearm flexors, levator scapulae, and triceps brachii. When the patient stands up, their palm, forearm, and back contact the palm of the robot's robotic arm, and the electrodes directly sense surface electromyographic potentials. The acquisition process uses an analog-to-digital converter with a 16-bit resolution and records the EMG signals at a sampling frequency of 1000Hz. Before entering the cache, the signal is processed through a bandpass filter with a cutoff frequency lower limit of 0.5Hz and an upper limit of 200Hz. A notch filter is also configured to eliminate 50Hz power-frequency noise. The processed sequence is timestamped with millisecond precision, derived from the camera's internal clock. Subsequently, the EMG data is synchronized with the posture data based on the frame number of the spatial trajectory captured by the camera. For example, at time 3.25s, the camera records the torso center of mass position as (0.38m, 0.22m, 0.90m). Simultaneously, the output voltage amplitude of the forearm EMG electrode is 0.54V, and the output voltage amplitude of the scapula electrode is 0.72V. The data at this time point is correlated into a one-to-one posture-tension data unit, ultimately forming complete posture-correlated muscle tension change data.

[0010] Step S2: analyzing the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data, and then performing center of gravity offset regression deconstruction to obtain a center of gravity offset regression trend; generating robot-assisted output parameters based on the center of gravity offset regression trend; In this embodiment of the present invention, after generating the posture-tension correlation data, the spatial coordinates of the patient's rising process must first be reconstructed. This reconstruction uses a human body key point extraction method to extract the coordinates of the hip, knee, ankle, shoulder, elbow, and lumbar spine center points from the 3D point cloud. For example, in the 25th frame of data, the coordinates of the left hip joint are (0.22m, 0.16m, 0.85m), the right hip joint is (0.21m, 0.15m, 0.85m), and the center of mass position is (0.215m, 0.155m, 0.87m). Simultaneously, the muscle electromyographic tension data is processed using first-order differences to obtain the muscle contraction rate. A rate exceeding 0.05V / s is considered to have produced a significant change in tension for that muscle. In this way, multiple sets of tension vectors are obtained during a specific rising phase. Next, with the vertical direction of the trunk as the reference axis, the tension change of each muscle group is decomposed into three components: the anterior-posterior direction (x-axis), the left-right direction (y-axis), and the vertical direction (z-axis). The resulting component data constitute a set of spatial force line distribution vectors. Subsequently, an orthogonal second-order decomposition is performed on the vector set to obtain the variance in each direction, thereby calculating the degree of imbalance of the muscle group force lines in different directions. For example, when the patient stood up for 5.50 seconds, the tension in the left upper arm in the y direction was 0.85N, and the tension in the right upper arm in the y direction was 1.25N. The difference was 0.40N. By comparing it with the mean square error of all data, the intensity of the force line imbalance was quantified as 35%. Next, the force line imbalance intensity was mapped to the dynamic trajectory of the center of mass. The center of mass trajectory was calculated through a frame sequence. From crawling to fully upright, the center of mass rose by a total of 0.30m. The displacement curve of the entire rising process was fitted with a third-order polynomial to form a regression trend of the center of gravity trajectory of the body. The trend was differentiated to obtain the changes in the center of gravity offset in the horizontal and vertical directions. For example, in the experiment, the lateral offset was 0.025m and the longitudinal offset was 0.30m. The overall upward trend was stable, but the lateral offset coupled the force line imbalance intensity by 35%. This value was directly output as the key input information in the subsequent robot-assisted output parameters.

[0011] In another embodiment, the posture-associated muscle tension change data generated in step S1 is used to analyze the imbalance of muscle force and the offset relationship between the body's center of gravity to generate auxiliary output parameters that the robot needs to execute. First, the posture-associated muscle tension change data is analyzed for the strength of the muscle group force line imbalance. Since the electromyographic signal comes from the patient's back area where the robot's palm contacts, the analysis focuses on the coordinated force pattern of the lower erector spinae and trapezius muscles in this area. The tension intensity values ​​of the 16 electromyographic channels are constructed into a 4x4 two-dimensional matrix, representing the muscle tension distribution map on the sensor contact surface. At each time point, the matrix is ​​weighted averaged to obtain the coordinates of the center of mass of the muscle activity intensity at that moment on the two-dimensional plane. The degree of deviation of this center of mass coordinate from the geometric center of the sensor is defined as the force line imbalance vector of the local muscle group. The modulus of this vector is the force line imbalance intensity. Subsequently, the center of gravity offset regression deconstruction is performed. Using the 3D coordinates of the 21 joints obtained in step S1 and the standard human segment mass distribution parameters, the 3D center of gravity coordinates of the patient's entire body are calculated for each frame through a weighted summation method, thereby constructing a continuous dynamic center of gravity trajectory. Next, the intensity of the force imbalance at each moment is correlated with the horizontal velocity of the dynamic center of gravity at that moment. Specifically, a univariate linear regression analysis is performed using the "force imbalance intensity" in the continuous time series as the dependent variable and the "horizontal velocity of the center of gravity" as the independent variable. The goal of this analysis is to establish a quantitative relationship: the extent to which rapid center of gravity movement causes imbalance in the back support muscles. The regression analysis yields a regression coefficient, which represents the regression trend of the center of gravity shift. For example, a calculated regression coefficient of 0.08 indicates that for every 0.1 m / s increase in the patient's horizontal velocity, the active center of mass of the back muscles will deviate by 8 mm from the contact center. Finally, the robot-assisted output parameters are generated based on this regression trend. A trigger threshold is set. When the real-time calculated force imbalance exceeds 0.015m (1.5cm) and the horizontal velocity of the center of gravity exceeds 0.25m, the system determines that the patient is unstable. The robot-assisted output parameters generated at this time are: an auxiliary force vector, whose direction is opposite to the force imbalance vector, aimed at pushing the center of mass of muscle activity back to the contact center; the magnitude of the auxiliary force is calculated by a proportional-integral controller, and its set value is 500 times the force imbalance. For example, when the imbalance is 0.016m, the auxiliary force is 8N; the auxiliary force is applied at the current contact point of the robot.

[0012] Step S3: The robot-assisted output parameters are used as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture, and the architecture is sent to the control terminal of the nursing robot to execute the control of the nursing robot.

[0013] In an embodiment of the present invention, the auxiliary output parameters obtained in step S2 are composed of two parts, one is the center of gravity offset regression trend sequence, and the other is the associated muscle group force line imbalance strength value. These parameters are input to the central control unit of the nursing robot. The central control unit first performs parameter feature weight distribution. According to the preset rules, the weight of the center of gravity offset trend is set to 0.6, and the weight of the muscle group force line imbalance strength is set to 0.4. After feature learning, an auxiliary parameter instruction set is generated. The auxiliary parameter instruction set is divided into three elements: functional factors, timing factors and constraint factors. The functional factors include the balance support, power support and lateral stabilization of the robot manipulator; the timing factor requires that the parameter instruction be refreshed every 0.02s; the constraint factor limits the single output force to no more than 30N, and the angular velocity of the manipulator arm is maintained at no more than 0.05rad / s. Taking an actual experiment as an example, when a patient's center of gravity shifts 0.025m to the left and 0.30m longitudinally while rising, the robot's right arm actuator generates a supporting force of 9N and its left arm actuator generates a support force of 6N according to auxiliary parameter instructions. The robot's palm contacts the patient's back and forearm, maintaining contact force within the specified range. The drive control frequency is maintained at 50Hz to ensure process continuity. The control architecture transmits each set of multi-dimensional instructions to the actuator end joint actuator via a data bus, achieving auxiliary support and dynamic coordination. In this way, the nursing robot's movements are precisely synchronized with the patient's motion trajectory throughout the rising process.

[0014] Step S1 includes the following steps: Step S11: using a 3D camera to capture the patient's posture trajectory when he or she stands up during the nursing process; Step S12: Using an electromyographic signal sensor and based on the posture trajectory, collecting muscle tension change signals of the patient's body contact areas when the patient stands up during the nursing process; the body contact areas include: the hand / forearm area, the back and shoulder blade area, and the elbow / upper arm area; Step S13: filtering the muscle tension change signal for noise, and then embedding the timestamp to obtain a change signal time series; Step S14: performing synchronous correlation processing on the time series of the change signal based on the posture trajectory to obtain posture-related muscle tension change data.

[0015] In this embodiment of the present invention, a 3D camera is fixed to the head of the nursing robot. The camera uses structured light to construct a spatial point cloud with a frame rate of 30Hz, outputting 200,000 spatial points per frame. The acquisition range is fixed to a cubic space of 1.2m long, 0.8m wide, and 1.5m high, covering the entire range of motion from sitting to standing. During the experiment, a patient was placed at the edge of a nursing bed and moved from a lying position to a sitting position and then gradually stood up. The entire process took approximately 6.5 seconds, and the camera was able to continuously capture 195 frames of motion data. The human skeleton key point recognition algorithm was invoked from the point cloud data to extract the coordinates of 29 joints, including the hip, knee, ankle, shoulder, elbow, and lumbar vertebrae center. The coordinates of each joint are expressed in meters (m) in a 3D Euclidean coordinate system. For example, at time 2.50s, the center of mass coordinates are (0.38m, 0.20m, 0.92m), the left knee joint coordinates are (0.20m, 0.32m, 0.53m), the right knee joint coordinates are (0.21m, 0.33m, 0.52m), and the torso angle is 15°. Successive frames form a posture trajectory data sequence. The trajectory includes not only the joint position coordinates but also the rate of change of each joint's angle relative to the vertical axis. This trajectory data provides a spatial reference for subsequent temporal matching of muscle tension data.

[0016] A four-channel surface electromyography (SEM) sensor array was embedded in the palm region of the nursing robot's robotic arm. Each channel consisted of two 10mm-diameter solid-state electrodes with a 20mm inter-electrode spacing, a sampling frequency of 1000Hz, and a resolution of 16 bits. The sensor array was affixed to the robot's contact surfaces at the patient's palm, forearm, elbow, and scapula. During the experiment, as the patient stood up and grasped the robotic arm, the hand electrodes measured flexor muscle group potential signals. For example, during the patient's 2.80s rise, the hand electrodes measured a peak voltage of 0.55V, an EMG amplitude of 0.48V at the elbow, and an amplitude of 0.67V at the scapular muscles. The collected data was a voltage-time series in volt-seconds (V·s). To ensure alignment with the posture trajectory, each acquisition included a precise sampling timestamp with an error of less than 1ms. This EMG signal reflects the tension and intensity of each contact muscle group during movement, providing the fundamental input for force line direction analysis and center of gravity offset estimation.

[0017] After data acquisition is complete, the EMG signals are preprocessed. This preprocessing utilizes a bandpass filter with a lower limit of 0.5 Hz and an upper limit of 200 Hz to remove baseline drift and high-frequency interference. To eliminate 50 Hz interference from the power supply system, an additional notch filter (bandwidth ±2 Hz) is added for suppression. The filtered signal undergoes further rectification and envelope extraction to reveal the temporal evolution of EMG energy. In the experiment, from the time the patient stood up to 4.10 s, the amplitude of the filtered envelope signal at the scapular electrode remained at 0.75 V, with a noise level less than 0.05 V, achieving a signal-to-noise ratio of 15:1. The processed signal data stream is timestamped at millisecond intervals using a unified clock. This timestamp is generated using the clock pulses shared by the camera and sensor, ensuring that all data sequences are aligned with the same time base. The final output is a sequence of "potential intensity-time" pairs. For example, at 4.00s, the hand signal amplitude is 0.48V, corresponding to the timestamp of 4000ms, and the back signal amplitude is 0.80V, corresponding to the timestamp of 4000ms. This sequence lays the foundation for synchronous processing.

[0018] Using a timestamp synchronization mechanism, the posture trajectory data sampled from the 3D camera is aligned millisecond-by-millisecond with the processed EMG time series. For each timestamp, the corresponding center of mass coordinates and local joint data are found and bound to the corresponding electrode potential amplitude. For example, at 3.50 seconds, the center of mass position is (0.36m, 0.24m, 0.95m), while the forearm electrode voltage is 0.60V and the scapula electrode voltage is 0.73V. The posture-tension correlation data at this moment is stored in a joint sequence. Through continuous synchronization, a complete posture-related muscle tension change data sequence is obtained. This dataset includes the spatial coordinates, joint angles, center of mass offset, and EMG potential amplitude at each time point. During further processing, the data is interpolated to eliminate data gaps caused by the difference between the camera frame rate and the EMG sensor sampling rate. The resulting posture-tension joint dataset covers the entire 6.5-second motion process with an accuracy of 1ms and an alignment error of less than 0.002s. This implementation process lays the foundation for the one-to-one correspondence of data in time and space for the subsequent force line vector decomposition and center of gravity regression.

[0019] Step S2 includes the following steps: Step S21: constructing the spatial coordinates of the posture trajectory during the standing up process to obtain the spatial coordinates of the posture trajectory; Step S22: analyzing the muscle group force line imbalance strength during the posture change process on the posture-related muscle tension change data, thereby obtaining the force line imbalance strength; Step S23: performing gravity center offset regression deconstruction on the force line imbalance intensity based on the posture trajectory spatial coordinates to obtain a gravity center offset regression trend; Step S24: Generate robot-assisted output parameters based on the center of gravity offset regression trend, combined with the posture trajectory space coordinates and posture-related muscle tension change data.

[0020] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: constructing the spatial coordinates of the posture trajectory during the standing up process to obtain the spatial coordinates of the posture trajectory; In an embodiment of the present invention, the patient posture trajectory established in step S1 is used to construct its spatial coordinates. The three-dimensional camera outputs point cloud data in a continuous time series and collects 30 frames per second. The three-dimensional coordinates of the hip joint, knee joint, ankle joint, shoulder joint, elbow joint and lumbar vertebrae center point are determined by the human joint detection algorithm for each frame of point cloud. For example, at 2.00s, the coordinates of the left hip joint are (0.22m, 0.18m, 0.82m), the right hip joint is (0.20m, 0.17m, 0.81m), and the center of mass coordinates are (0.21m, 0.175m, 0.835m). The spatial coordinate construction is not only a single coordinate record, but also forms a continuous trajectory curve through a sequence of joint points. During the process of getting up, the speed and angle changes of each joint are also calculated by differential calculation of adjacent frames. Taking the right knee joint as an example, at 2.00s, the coordinates are (0.18m, 0.30m, 0.50m), and at 2.03s, the coordinates are (0.18m, 0.305m, 0.52m). The difference between the two points outputs a local velocity vector of (0.0m / s, 0.17m / s, 0.67m / s). By performing this operation on all joints, the dynamic coordinate matrix for the entire upper and lower body joints is obtained. This dynamic matrix contains not only the absolute position but also the rate of change information, ultimately forming a complete spatial coordinate sequence for the standing trajectory, providing precise spatial support for subsequent mechanical analysis.

[0021] Step S22: analyzing the muscle group force line imbalance strength during the posture change process on the posture-related muscle tension change data, thereby obtaining the force line imbalance strength; In an embodiment of the present invention, after spatial coordinates are established, force imbalance intensity analysis is performed based on the synchronized sequence of posture and muscle tension data. The collected electromyographic signals are filtered and converted into muscle group tension amplitudes. The electrical signals, measured in volts, are linearly related to muscle tension. Through experimental calibration, the electrical signal amplitudes are mapped to equivalent force values. For example, the voltage amplitude of the forearm electrode at 3.00s is 0.55V, which corresponds to a force value of 10.2N. The force vector of each muscle group in the contact area is decomposed according to the body's three-dimensional coordinate system, yielding components along the x, y, and z directions. For example, the force corresponding to the back muscles at 3.20s is 15N, with the decomposition results being 4.0N in the x-direction, 3.5N in the y-direction, and 7.5N in the z-direction. The data for the left and right symmetrical muscle groups are then subtracted. When the difference between the two symmetrical areas exceeds a set threshold of 5N, a force imbalance is defined. For example, the difference in the y-direction between the left and right arms is 6.2N, which exceeds the threshold and is therefore considered an imbalance. Throughout the entire standing-up movement, the imbalance intensity data, collected frame by frame, forms a time curve. For example, between 3.00s and 4.50s, the maximum imbalance intensity is 35%. This process measures the mechanical heterogeneity of muscle groups during postural changes, and the resulting time series serves as input for subsequent center of gravity trend regression.

[0022] Step S23: performing gravity center offset regression deconstruction on the force line imbalance intensity based on the posture trajectory spatial coordinates to obtain a gravity center offset regression trend; In an embodiment of the present invention, spatial coordinates are used to calculate the patient's center of mass trajectory. The center of mass position is obtained by weighted averaging of key points throughout the body. During the process of standing up, the center of mass rises from the coordinates (0.21m, 0.175m, 0.835m) to (0.22m, 0.23m, 1.12m), with a longitudinal offset of 0.285m and a lateral offset of 0.01m. The center of mass displacement curve is fitted with a third-order polynomial, and the fitting curve represents the continuous change trend of the center of gravity. In order to accurately reflect the impact of muscle strength imbalance, the force line imbalance intensity sequence and the center of gravity trajectory are coupled for analysis. At each moment, the imbalance intensity is mapped to the corresponding center of mass coordinate point to form a "local imbalance-center of mass position" binary relationship. For example, at 3.50s, the imbalance intensity is 28% and the center of mass position is (0.22m, 0.20m, 0.98m). This mapping relationship is then differentiated to calculate the slight offset trend of the center of gravity with the imbalance intensity, thereby deconstructing the center of gravity offset regression trend. During the experiment, the longitudinal acceleration measured between 4.00s and 4.30s was 0.20m / s², corresponding to a lateral imbalance that resulted in a lateral center of gravity shift of 0.025m. This value serves as the regression decomposition result for the center of gravity shift, providing a quantitative basis for the subsequent auxiliary parameter generation.

[0023] In another embodiment, a quantitative relationship between the strength of the force line imbalance and the dynamic change of the patient's body center of gravity is established to deconstruct the center of gravity offset regression trend. First, based on the posture trajectory spatial coordinates obtained in step S21, the overall three-dimensional center of gravity coordinates of the patient's body are calculated for each frame. The calculation is done using an anthropometric parameter set containing 15 body segments (such as the head, trunk, left and right thighs, calves, etc.). This parameter set defines for each segment the percentage of its mass to the total body weight (for example, 49.7% for the trunk) and the relative position of its center of mass within the segment. The center of mass coordinates of the whole body are obtained by taking the mass-weighted average of the center of mass coordinates of all segments. Next, the displacement velocity of the center of mass in the horizontal plane (XY plane) is calculated. , calculated using the backward difference method: ,in is the time interval between two frames (here 1 / 30 second). Then, the horizontal displacement velocity sequence of the center of gravity during the entire standing-up process is paired with the force line imbalance intensity sequence obtained in step S22. The two sets of data are subjected to a univariate linear regression analysis using the least squares method to find the best fitting line. Here, is the slope of the regression line, is the intercept. The slope It is defined as the center of gravity shift regression trend. For example, if the calculation is =0.05s, which means in physics that for every 1m / s increase in the velocity of the patient's center of gravity on the horizontal plane, the imbalance strength of the back muscle force line is expected to increase by 0.05m. It is output as a key parameter for subsequent robot-assisted decision-making.

[0024] Step S24: Generate robot-assisted output parameters based on the center of gravity offset regression trend, combined with the posture trajectory space coordinates and posture-related muscle tension change data.

[0025] In this embodiment of the present invention, after obtaining the center of gravity offset regression trend, it needs to be converted into auxiliary output parameters for the nursing robot to perform actions. This process is performed in conjunction with the spatial coordinates of the posture trajectory and the posture-tension data. First, the offset direction and magnitude are determined based on the regression trend, for example, a lateral offset of 0.025m and a longitudinal displacement of 0.285m. Second, the amount of support force to be applied by the robot is determined based on the muscle tension margin. In experiments, when the scapular muscle group is imbalanced and insufficiently loaded, the robot arm is required to provide an additional 9N of support force. The support force calculation process uses the difference between the regression trend offset and the electromyographic tension data to obtain the missing compensation value. Finally, the compensation force is used as a parameter and output synchronously with the motion trajectory of the robot arm end. For example, from the time of standing up to 4.20s, the auxiliary output parameters include: right arm output support force of 9N, left arm output support force of 6N, and hand contact angle adjustment of 3°. These parameters are transmitted to the control terminal in real time to ensure that the robot can provide continuous and stable assistance despite the patient's center of mass rising, lateral offset, and muscle force imbalance.

[0026] In another embodiment, an auxiliary trigger condition is set: when the real-time monitored force line imbalance intensity When the preset safety threshold of 0.03 meters is exceeded, the system immediately generates auxiliary parameters. The generated parameters include three parts: the point of action, direction and magnitude of the auxiliary force. The point of action of the auxiliary force is set to the current contact point between the end of the robot's manipulator and the patient's body. The three-dimensional coordinates of this point are obtained by the real-time forward kinematics solution of the robot. The direction of the auxiliary force is designed to be opposite to the direction of the force line imbalance to play a corrective role. Specifically, in the local two-dimensional coordinate system of the robot's palm, the direction vector of the force line imbalance is , so the direction vector of the auxiliary force is This two-dimensional vector will be transformed into the world coordinate system through the posture matrix of the robot end, becoming a three-dimensional force direction vector. This is calculated through a proportional controller, whose magnitude is proportional to the amount by which the force imbalance intensity exceeds the threshold, and is adjusted in conjunction with the center of gravity offset regression trend: Here, the proportional gain Is a preset constant, such as 200N / m. For example, when the real-time force line imbalance intensity is 0.035m, the horizontal speed of the center of gravity is 0.2m / s, and the regression trend is When the speed is 0.05 seconds per meter, the calculated auxiliary force is (N). Finally, this set of robot-assisted output parameters including the three-dimensional coordinates of the action point, the three-dimensional force direction vector, and the force magnitude scalar is packaged and sent to the robot control system.

[0027] Step S22 includes the following steps: Step S221: calculating the average difference in tension intensity changes during posture changes based on the posture-associated muscle tension change data; Step S222: performing a three-dimensional force line direction vector decomposition of the posture-related muscle tension change data according to the tension intensity change mean difference during the posture change process to obtain a spatial force line distribution vector set; Step S223: performing orthogonal second-order statistical decomposition on the spatial force line distribution vector set to obtain force line second-order decomposition data; Step S224: analyzing the imbalance strength of muscle group force lines during posture changes using the second-order force line decomposition data on the spatial force line distribution vector set, thereby obtaining the force line imbalance strength.

[0028] As an example of the present invention, refer to Figure 3 As shown, in this example, step S22 includes: Step S221: calculating the average difference in tension intensity changes during posture changes based on the posture-associated muscle tension change data; In an embodiment of the present invention, the collected posture-related muscle tension change data are centrally organized, and the myoelectric signal amplitudes of the forearm, upper arm, and scapular back at each time point are uniformly expressed with their corresponding coordinate information. For example, in the experiment, the sampling period is set to 1ms, the total sampling time is 6.5s, and the total data points are obtained to be 6500. The myoelectric signal voltage amplitude is mapped to the muscle equivalent value after filtering, and the mapping relationship is established by the experimental calibration curve, such as a potential amplitude of 0.50V corresponding to a tension of 10N, and a potential amplitude of 0.70V corresponding to a tension of 14N. Subsequently, a weighted average is taken for the tension values ​​of multiple contact areas in the same time series to obtain the comprehensive tension strength of a single frame. The mean difference in tension strength change is then obtained by calculating the average deviation of the difference between consecutive frames. Taking real experimental data as an example, between 3.20s and 3.25s, forearm tension increased from 10.2N to 11.6N, back tension from 14.0N to 15.5N, and scapular tension from 12.2N to 12.8N. Summing and averaging these three changes yielded a mean difference of +1.7N for this period. Continuously calculating the mean difference across all sampling intervals yielded a complete sequence of mean differences in tension intensity. This sequence reflects the overall fluctuations in force applied to different muscle groups during standing up, providing a foundation for subsequent three-dimensional vector decomposition.

[0029] In another embodiment, based on the posture-associated muscle tension change data, the short-term volatility of muscle tension intensity during the posture change process is calculated and quantified as the mean difference in tension intensity change. The calculation is performed within a sliding time window with a length of 6 consecutive time points (i.e., 0.2 seconds). For each time point, the system first extracts the tension intensity data of the three core muscle groups from the posture-associated muscle tension change data. These three muscle groups correspond to the "back and scapular area", "hand / forearm area" and "elbow / upper arm area" collected in step S12, respectively. The tension intensity value of each area is obtained by taking the arithmetic average of the root mean square voltage values ​​of the corresponding multiple electromyographic signal channels. For example, the tension intensity value of the "back and scapular area" is It is obtained by adding the RMS voltage values ​​of the corresponding 4 channels and dividing by 4. Then, within the sliding window, the first-order backward difference of these 6 consecutive tension strength values ​​is calculated to obtain 5 consecutive tension strength change values: ,in Traverse from the second point to the sixth point in the window. Finally, calculate the arithmetic mean of the absolute values ​​of these five change values, and the result is the mean difference of the tension intensity change corresponding to the current time point. , and its calculation process is: This sliding window calculation is repeated for each time point in the entire standing-up process, thereby generating a synchronized time series of the mean difference in tension intensity for each muscle group, which characterizes the intensity of its tension fluctuation.

[0030] Step S222: performing a three-dimensional force line direction vector decomposition of the posture-related muscle tension change data according to the tension intensity change mean difference during the posture change process to obtain a spatial force line distribution vector set; In an embodiment of the present invention, after completing the calculation of the mean difference in tension strength change, it is necessary to project it into three-dimensional space for directional decomposition. First, clarify the coordinate reference system, select the coordinate system z-axis as the vertical direction, the y-axis as the front-to-back direction, and the x-axis as the left-right direction. Then project the mean difference in tension strength of each muscle group into the three-dimensional coordinate system. Taking the 3.40s moment in the experiment as an example, the mean difference in tension change of the left forearm is +1.4N. The main force direction of the muscle is calculated by the joint posture trajectory to be 20° with the x-axis, 45° with the y-axis, and 55° with the z-axis. The three directional components thus decomposed are 0.48N in the x-direction, 0.99N in the y-direction, and 0.80N in the z-direction. The same method is applied to the back and scapular region, and at the same time, the tension mean difference x component is -0.35N, the y component is 1.20N, and the z component is 1.55N. In this way, the mean difference in tension change of all muscle groups is decomposed into vector form to form a spatial force line distribution vector. The entire standing-up process is continuously calculated at each time point, generating a spatial force distribution vector set—a set of multidimensional vector data that varies over time. This vector set clearly describes the direction of force contribution and its trend over time, serving as the mathematical input for further orthogonal decomposition.

[0031] Step S223: performing orthogonal second-order statistical decomposition on the spatial force line distribution vector set to obtain force line second-order decomposition data; In an embodiment of the present invention, after obtaining the spatial force line distribution vector set, it is then decomposed by an orthogonal second-order statistical method. The purpose of the decomposition is to eliminate the linear coupling relationship between the directions and extract independent main variations. The specific process is: the mean and variance of the three-direction components are calculated for the vector set in each time period, and then the covariance matrix between different directions is obtained. Taking the experimental data as an example, during the period from 3.60s to 3.70s, the mean in the x-direction is 0.52N and the variance is 0.04(N²), the mean in the y-direction is 1.10N and the variance is 0.06(N²), and the mean in the z-direction is 1.20N and the variance is 0.05(N²). In the covariance matrix, the covariance in the xy direction is 0.015, the covariance in the yz direction is 0.018, and the covariance in the xz direction is 0.012. Through orthogonal decomposition, the matrix is ​​converted into three independent principal components, which are called second-order decomposition vectors. The decomposition results show that during this time period, the first principal component, along the yz plane, contributed 62%, corresponding to a change in intensity of 1.45 N. The second principal component, along the 40° angle between the x and z axes, contributed 25%, corresponding to an intensity of 0.85 N. The third principal component contributed 13%, corresponding to an intensity of 0.36 N. By decomposing the force line by time period, a second-order decomposition data set was formed.

[0032] Step S224: analyzing the imbalance strength of muscle group force lines during posture changes using the second-order force line decomposition data on the spatial force line distribution vector set, thereby obtaining the force line imbalance strength.

[0033] In an embodiment of the present invention, after obtaining the second-order decomposition data of the force line, it is used to analyze the imbalance strength of the force lines of different muscle groups during posture changes. The specific implementation process is: in each time period, the force lines of the left and right symmetrical muscle groups and the front and back muscle groups are compared, and the average result of the square of the difference is used as the quantitative value of the degree of imbalance. Taking the 4.00s moment in the experiment as an example, the second-order principal component strength of the front direction of the left arm is 1.20N, and the principal component strength of the corresponding direction of the right arm is 1.75N, and the difference between the two is 0.55N; the strength of the back muscle group and the scapular muscle group in the vertical direction is 1.45N and 2.10N respectively, with a difference of 0.65N. By standardizing all the differences, the imbalance strength index is defined as the square root of the sum of the squares of each difference. The result of this calculation at 4.00s is 0.85N. After normalizing this result with the parameters related to body mass, it is concluded that the imbalance strength accounts for 32%. By calculating the imbalance intensity at each moment of the entire movement process, the complete imbalance intensity can be generated. For example, the maximum value occurs in the range of 3.80s to 4.20s, and the maximum imbalance intensity is 38%.

[0034] Step S23 includes the following steps: Step S231: Deconstructing the spatial coordinates of the posture trajectory to obtain the speed change rate, joint motion angles, and trunk posture angles when the patient stands up during the nursing process; the joints include the hip joint, knee joint, ankle joint, shoulder joint, and elbow joint; and the trunk refers to the area from the pelvis to the shoulder. Step S232: obtaining a dynamic center of gravity trajectory based on the velocity change rate, joint motion angle, and trunk posture angle; performing a center of gravity displacement vector polynomial fitting on the dynamic center of gravity trajectory to generate a center of gravity displacement fitting vector; Step S233: performing coordinate mapping processing on the force line imbalance intensity based on the posture trajectory spatial coordinates to generate local force line imbalance distribution data corresponding to each spatial coordinate point; Step S234: performing gravity center trend vector differentiation on the local force line imbalance distribution data based on the gravity center displacement fitting vector to obtain a gravity center offset trend sequence; Step S235: performing correlation regression analysis on the center of gravity shift trend sequence, and then performing center of gravity shift regression deconstruction to obtain the center of gravity shift regression trend.

[0035] In an embodiment of the present invention, the collected spatial coordinates of the posture trajectory are deconstructed frame by frame. The three-dimensional camera outputs the point cloud data of the patient's getting up process at a frame rate of 30Hz, and the joint recognition module extracts the spatial coordinates of the hip joint, knee joint, ankle joint, shoulder joint, elbow joint and lumbar center point. In order to obtain the velocity change rate, it is necessary to make a difference between consecutive frames. For example, when the patient is at 2.00s, the coordinates of the left knee joint are (0.20m, 0.32m, 0.51m), and at 2.03s, the coordinates of the point are (0.20m, 0.325m, 0.53m). The difference is divided by the time interval of 0.03s, and the velocity change rate of the left knee joint is (0.0m / s, 0.17m / s, 0.67m / s). In the joint angle calculation, the angle between the two limb vectors connected by the joint is solved. For example, the hip joint is formed by the thigh vector and the trunk vector. During the experiment, the left hip joint angle was 102° at 3.00 seconds, gradually extending to 174° at 3.80 seconds. Simultaneously, the trunk posture angle was calculated by the angle between the line connecting the pelvis and the midpoint of the shoulder and the vertical axis. In the test case, the trunk forward tilt angle was approximately 25° at 2.50 seconds, decreasing to 8° at 4.20 seconds. These calculations yielded a complete time series data matrix of velocity change rates, joint motion angles, and trunk posture angles, clearly describing the body's dynamic characteristics.

[0036] After completing the joint and trunk deconstruction, the dynamic center of mass trajectory during the rise process needs to be calculated. The center of mass position is obtained by the weighted average of the hip joints, knee joints, ankle joints, shoulder joints, and the midpoint of the trunk. The weighting coefficient is set according to the proportion of each part in the total body mass. For example, the weight of the two hip joints and the lumbar center point is 0.40, the weight of the two knee joints is 0.25, the weight of the two shoulders is 0.20, and the weight of the two ankle joints is 0.15. During the experiment, the patient's center of mass initially positioned at (0.215m, 0.175m, 0.835m). As the movement gradually shifted, it reached (0.225m, 0.240m, 1.120m) at 4.80s, with a longitudinal displacement of approximately 0.285m and a lateral displacement of approximately 0.010m. The collected center of mass trajectory was interpolated to form a continuous sequence over time. A third-order polynomial fit was then performed on the center of mass trajectory, and the fitting expression is: ,in Indicates the time to stand up, in seconds (s). is the initial center of gravity position component, is the coefficient of the center of gravity velocity term, reflecting a linear change, is the coefficient of the center of gravity acceleration term, is a high-order nonlinear coefficient. Taking the longitudinal component of the experimental center of gravity as an example, the coefficient of determination of the fitting result is 0.98, indicating that the fitting curve accurately reflects the true trajectory. The resulting center of gravity displacement fitting vector is composed of continuous fitting terms in the x, y, and z directions, providing a mathematical foundation for subsequent trend analysis.

[0037] After obtaining the dynamic center of gravity fitting trajectory, it is necessary to establish a mapping relationship between the force line imbalance intensity data and the posture trajectory space coordinates. First, the force line imbalance intensity comes from the calculation result of step S22, which changes with time in units of N. Secondly, according to the position of the posture trajectory in the three-dimensional coordinates, the local space point at each moment is determined. For example, at 3.60s, the center of mass coordinates are (0.220m, 0.205m, 0.980m), and the force line imbalance intensity at this time is 0.82N. Bind the two by timestamp and assign the intensity to the corresponding center of mass position. Perform this operation for all moments to form a mapping set of "coordinate point-imbalance intensity". In order to further obtain the local distribution, the joint coordinates are also included in the mapping range. For example, at 3.70s, the right knee joint coordinates (0.185m, 0.315m, 0.540m) were assigned an imbalance strength component of 0.65N, and the left shoulder joint coordinates (0.245m, 0.210m, 1.200m) were assigned a value of 0.95N. After processing each joint and center of mass point, local imbalance distribution data covering key points throughout the body was generated. This data not only includes spatial position but also describes the current force imbalance state. Subsequent regression deconstruction can reveal the corresponding relationship between center of gravity trends and muscle group force imbalances, providing the necessary and precise data for further generation of robot output parameters.

[0038] The center of gravity displacement fitting vector obtained in step S232 is a three-dimensional function that changes with time and reflects the displacement change trends in the x-direction, y-direction and z-direction respectively. Secondly, the local force line imbalance distribution data generated in step S233 is used to synchronize the imbalance intensity value of each joint and center of mass position with the corresponding spatial coordinates. In order to calculate the center of gravity trend difference, the correspondence between the center of mass fitting vector and the imbalance intensity of the position is first determined at the same time point. For example, at 3.80s, the center of mass coordinates are (0.222m, 0.218m, 1.015m), and the fitting vector outputs a longitudinal displacement speed of 0.25m / s, corresponding to a local imbalance intensity of 0.82N. Taking this moment as a reference, the data of adjacent frames are differentiated to obtain the differential vector under the joint action of the center of gravity displacement change and the local imbalance intensity. The specific method is: subtract the fitting vector between time points t and t+Δt, and then combine it with the imbalance value to form a correction vector. For example, between 3.80s and 3.83s, the longitudinal displacement increased by 0.012m and the lateral displacement increased by 0.002m, corresponding to an increase of 0.20N in the imbalance correction component. This results in a differential trend in this interval that shows an "increase with an increase in leftward excursion." By repeating the differential operation over the entire time series, a center of gravity shift trend sequence is ultimately formed. This sequence, with time as the primary axis, records the displacement direction and rate of the center of gravity in three-dimensional space. Coupled with the imbalance intensity, it accurately reflects the dynamic characteristics of the center of gravity at each stage of the patient's rising process.

[0039] After the trend sequence is generated, regression analysis is performed to extract the overall pattern. Specifically, the center of gravity shift trend sequence is divided into three-dimensional components: the horizontal component x(t), the longitudinal component y(t), and the vertical component z(t). For each component, regression is performed using the correlation between the time series and the imbalance intensity. Taking the longitudinal component as an example, its time-varying function is defined as: ,in, Indicates the longitudinal center of gravity shift trend, Indicates time, Represents the imbalance intensity series at time The value of is in N. is the initial longitudinal offset position, is the longitudinal velocity regression coefficient, is the acceleration regression coefficient, is the imbalance coupling coefficient. Taking the experimental data as an example, during the period from 2.50s to 4.50s, the total longitudinal upward displacement was approximately 0.285m, the maximum imbalance intensity was 0.95N, and the corresponding longitudinal regression expression determination coefficient was 0.96. The results of the lateral offset showed that the maximum offset was 0.025m, and the regression analysis showed that the imbalance term accounted for 38% of the trend contribution. The deconstruction process obtained a clear offset trend structure by splitting the regression function into a velocity trend part, an acceleration trend part, and an imbalance coupling part. For example, at 4.00s, the longitudinal trend was dominated by the velocity term, and the lateral trend was dominated by the imbalance part, indicating that the center of gravity rise was stable during this period but the lateral offset was significant. The final center of gravity offset regression trend is a multidimensional time function containing three components, which is used to guide the generation of auxiliary output parameters of the nursing robot.

[0040] Performing gravity center trend vector differentiation on the local force line imbalance distribution data includes the following steps: Performing a time-space coordinate cross-match based on the center-of-gravity displacement fitting vector and the local force line imbalance distribution data to obtain the center-of-gravity-force line coupling mapping data corresponding to each time series coordinate point; Performing differential vector field construction processing on adjacent time series coordinate points according to the center of gravity-force line coupling mapping data to obtain a local inhomogeneity differential vector field of a continuous time series; Performing gradient difference processing on each spatial coordinate point of the local inhomogeneity differential vector field to obtain a local center of gravity offset gradient sequence; Multi-scale trend coupling is performed on the local center of gravity offset gradient series, and the offset change rates at different time scales are combined to obtain the multi-scale center of gravity offset trend; Based on the multi-scale center of gravity shift trend, the center of gravity trend vector difference is performed to obtain the center of gravity shift trend sequence.

[0041] In this embodiment of the present invention, the center of mass displacement fitting vector obtained in step S232 is called, and the vector gives the center of mass position coordinates at each time point. , and the displacement trend in three directions. Meanwhile, the local force imbalance distribution data provided in step S233 includes the position of each joint And the force line imbalance intensity value F. In order to form a cross match, the two types of data need to be aligned at the same timestamp. For example, at 3.80s, the center of gravity fitting vector coordinates are (0.222m, 0.218m, 1.015m), and the longitudinal velocity is 0.25m / s; at this time, the right knee joint coordinates are (0.185m, 0.315m, 0.540m), the imbalance intensity value is 0.68N, and the left shoulder joint coordinates are (0.245m, 0.210m, 1.200m), and the imbalance intensity value is 0.92N. Through the intersection of time and space coordinates, the corresponding center of gravity vector is mapped to the local imbalance point of the joint. The specific operation is to calculate the difference between the center of gravity coordinates and the joint coordinates as a relative position vector, and then superimpose the imbalance intensity on it. For example, the relative center of gravity displacement vector of the right knee joint is (-0.037m, 0.097m, -0.475m), and the associated force imbalance intensity is 0.68N, ultimately generating the coupling data points (-0.037, 0.097, -0.475; 0.68). This operation is performed for all joints and center of gravity positions, resulting in a complete "center of gravity-force line" coupling mapping dataset at each moment. This dataset contains both spatial positional relationships and mechanical inhomogeneity characteristics, serving as the basic input for the subsequent differential vector field construction.

[0042] After obtaining the spatiotemporal cross-mapping data, a continuous time series differential vector field needs to be constructed. First, the coupling mapping data for each two adjacent time points are differentiated. This differentiation process involves the simultaneous calculation of position changes and imbalance intensity changes. For example, the center of mass-right knee coupling data at 3.80s and 3.83s are (-0.037, 0.097, -0.475; 0.68) and (-0.040, 0.102, -0.462; 0.72), respectively. Subtracting the two yields a differential vector of (-0.003m, 0.005m, 0.013m; 0.04N). The first three results represent time series displacement differences, while the last term represents the imbalance intensity difference. In this way, the differential values ​​of all joints relative to the center of mass are calculated. Next, these differential values ​​are concatenated to form a complete vector field. Each vector element in this vector field represents both the displacement direction trend and the increase or decrease in mechanical inhomogeneity. For example, at 3.83 seconds, the left shoulder joint differential result is (0.003m, 0.004m, 0.015m; 0.06N), indicating that the shoulder's spatial movement relative to the center of gravity during this phase is anterior-upward, accompanied by increasing imbalance. Combining the differential vectors of all joints and the center of mass forms a continuous time-series local inhomogeneity differential vector field. This vector field continuously describes the coupling effects of the patient's center of gravity shift and local mechanical anomalies during the rising process in the temporal dimension, providing precise input data for subsequent trend analysis and regression decomposition.

[0043] First, the continuous time-series local inhomogeneity difference vector field generated in the previous step is called. This vector field contains the relative displacements and directions of multiple joints relative to the center of mass, as well as the corresponding force imbalance differential values. At each timestamp, the vector field uses three-dimensional coordinate points as nodes. For example, at the end of 3.80 seconds, the difference result for the right knee node is (-0.003m, 0.005m, 0.013m; 0.04N), and the difference result for the left shoulder node is (0.003m, 0.004m, 0.015m; 0.06N). To obtain local spatial gradient features, the gradient of the difference vectors of adjacent joint nodes is calculated in the three-dimensional coordinate domain. The specific method is: at each joint node, its three-dimensional difference vector is subtracted from the vectors of the surrounding adjacent nodes, and then divided by their spatial distance. For example, if the difference vector for the shoulder and elbow joints, separated by 0.12 meters, is (0.006m, -0.002m, 0.010m), the local gradient is (0.050, -0.017, 0.083), expressed as a dimensionless ratio. This approach allows similar calculations to be performed on all joints, forming a complete local gradient matrix at each timestamp. This matrix reflects the uneven vector change rates at different locations. By sequentially arranging the matrices for all timestamps, we obtain a sequence of local center-of-gravity offset gradients. This sequence clearly expresses the offset gradient trajectory of each local point during center-of-gravity motion, laying the foundation for subsequent cross-scale coupling.

[0044] The local center of mass shift gradient series only reflects data at a single temporal resolution. To capture the combined patterns of short-term jitter and long-term shifts throughout the patient's rise, multi-scale processing is required. Specifically, the time series is divided into three scales: a fine scale of 0.03s, corresponding to a 30Hz camera sampling rate; a mesoscale of 0.30s, corresponding to a window of 10 frames; and a macroscale of 1.50s, corresponding to a window of 50 frames. At each scale, the mean and rate of change of the gradient series are calculated. For example, within a 0.30s window, when the local gradient mean is (0.045, 0.020, 0.078) and the rate of change is (0.003, 0.002, 0.006), it indicates a significant, stable upward trend in the center of mass during this period. Subsequently, the data at different scales are weighted and synthesized. The weight distribution principle is 0.4 for the fine scale, 0.35 for the mesoscale, and 0.25 for the macroscale, to ensure that small, short-term perturbations do not obscure the overall trend. Combined with experimental data, between 3.00 and 4.50 seconds after the patient stood up, the fine-scale data showed multiple localized lateral shifts of up to 0.008 m, while the macro-scale trend showed a sustained vertical rise of 0.28 m. The combined results form a multi-scale center of gravity shift trend, which not only includes the overall displacement direction but also accurately captures the cumulative effects at different time scales, more accurately reflecting the actual dynamic process.

[0045] After obtaining the multi-scale center of gravity shift trends, further processing is performed to obtain the final center of gravity shift trend sequence. The specific process is as follows: First, three-dimensional component curves are extracted from the multi-scale trend. For example, the longitudinal trend rises from an initial value of 0.835m to 1.120m, the lateral trend fluctuates between 0.005m and 0.025m, and the overall offset in the fore-aft direction (y-axis) does not exceed 0.010m. These trend curves are then gradually differentiated to obtain trend vector differences. For example, between 3.50s and 3.53s, the longitudinal differential value is (+0.012m) and the lateral differential value is (+0.002m). The longitudinal differential corresponds to a gradient increment of +0.05, and the lateral differential corresponds to an increment of +0.08. The differential results reflect the instantaneous magnitude of the center of gravity trend. The differential results from all time periods are concatenated to form a complete center of gravity shift trend sequence. This sequence is represented as a set of three-dimensional time functions, recording the direction and magnitude of the center of gravity trend gain at each moment. For example, in the experimental data, the maximum longitudinal trend increment occurs between 3.95s and 3.98s, with a value of +0.018m, while the maximum lateral trend increment occurs between 4.05s and 4.08s, with a value of +0.004m. This processing generates a complete time-varying center of gravity offset trend sequence, providing input data for generating precise mechanical auxiliary parameters for the nursing robot.

[0046] Step S24 includes the following steps: Step S241: Determine the patient's posture orientation when standing up based on the center of gravity offset regression trend and the posture trajectory spatial coordinates; Step S242: performing fatigue load analysis on the posture-related muscle tension change data to determine the acceptable muscle force margin; Step S243: Optimizing multi-point contact mechanics based on the posture orientation and the muscle acceptable force margin to obtain dynamic contact force; Step S244: performing adaptive fuzzy inverse kinematics transformation of the robot based on the dynamic contact force and the posture orientation to obtain the robot joint space motion trajectory; Step S245: performing data parameter fusion based on the robot joint spatial motion trajectory and dynamic contact force to generate robot auxiliary output parameters.

[0047] In an embodiment of the present invention, the center of gravity offset regression trend sequence obtained in step S235 is called. The sequence includes dynamic displacement trends in the three coordinate directions of x, y, and z. For example, in the experimental record, the average displacement in the longitudinal direction (z) is 0.285m, the maximum displacement in the lateral direction (x) is 0.025m, and the offset in the front-back direction (y) does not exceed 0.010m. At the same time, the spatial coordinate sequence of the posture trajectory provides the angles of key joints. For example, during the process of standing up, the left hip joint angle increases from 102° to 172°, the right knee joint angle extends from 85° to 176°, and the trunk inclination angle gradually decreases from 25° to 8°. By synchronizing the offset trend with the joint angle, the overall orientation of the posture tendency can be determined. For example, in the period from 3.80s to 4.10s, the longitudinal trajectory shows an upright trend, while the lateral peak offset reaches 0.022m, and the shoulder deviates to the left by about 6° relative to the center of mass. The comprehensive judgment of its posture tendency orientation is "longitudinal rise, left offset". The final posture inclination vector consists of two components: the center of gravity direction [0.022m, 0.008m, 0.285m] and the posture angle [θhip = 172°, θknee = 176°, θtorso = 8°]. Using the posture-related muscle tension change data obtained in step S1, fatigue load analysis was performed. Specifically, the cumulative tension integrals of the muscle groups over different time periods were calculated and compared with physiological limits. During the experiment, the peak voltage of the electrical signal of the forearm muscles during the rise phase (total duration of 6.5s) reached 0.65V, corresponding to an equivalent tension of approximately 12.8N. The cumulative integration of the tension over time yielded a total stress value of 58N·s. The standard physiological endurance threshold is 72N·s, resulting in a remaining available margin of 14N·s. The analysis of the back muscles yielded a stress value of 130N·s with a margin of 40N·s; the corresponding margin for the scapular muscles was 18N·s. Comparing the three sets of data reveals that the back muscles have the largest load tolerance, while the scapular muscles have the smallest. This indicates that, in robot-assisted control, the shoulder support force must not exceed its tolerance of 18N, the forearm must not exceed 14N, and the back's allowable compensation range is within 40N. This conclusion provides a strict upper limit for multi-point contact optimization, ensuring that contact force control is scientific and within the patient's physical capabilities.

[0048] The posture inclination and orientation of S241 were combined with the acceptable force margins for different muscle groups in S242. First, the contact points were defined as the left forearm contact area, the right forearm contact area, and the back support area. During the 3.90-second period, the posture inclination was determined to be a leftward deviation of 0.022m and a longitudinal rise of 0.285m. At this point, the lateral damping force borne by the left forearm contact should be greater than that borne by the right to correct the deviation. Based on the margins, the maximum allowable force on the left arm is 14N, so an actual force of 9N is allocated; the right arm is allocated 6N to maintain balance. The back, as the longitudinal support area, bears a longitudinal compensation force of 12N, which is less than the maximum margin of 40N, ensuring safety. The optimization process utilizes a matrix balancing method, coupling the force outputs of the three contact areas with the center of gravity orientation to align the resultant force direction with the main longitudinal rise direction while offsetting lateral deviation. The final output is a set of dynamic contact force vectors: left forearm (9N, -0.002m), right forearm (6N, 0.003m), and back (12N, 0.000m). These parameters ensure overall balance and serve as precise force inputs for the robot's actuators.

[0049] The robot joint motion trajectory is constructed based on the dynamic contact force vectors and posture orientations obtained by S243. The nursing robot arm contains a 6-DOF joint structure: the shoulder joint (3 rotation axes), the elbow joint (1 rotation axis), and the wrist joint (2 rotation axes). The input parameters are the contact forces Fx, Fy, and Fz at the point of application and the target orientation angles θtorso, θhip, and θknee. Using inverse kinematics, the target position of the end effector at time t is mapped to a sequence of joint angles. For example, at 3.95 seconds, a back support force of 12N corresponds to a vertical support rise, requiring the end arm to rise 0.012m. The calculated results are a shoulder rotation angle of +0.06rad and an elbow extension angle of +0.04rad. Meanwhile, a leftward deviation in the posture corresponds to the end movement direction (0.004m, -0.002m), resulting in a wrist rotation angle of -0.03rad around the y-axis. A fuzzy control strategy ensures that the input dynamic contact force remains within a margin while still resolving a valid trajectory. This means that when the target and resolved joint angles do not fully match, the longitudinal support requirement is prioritized with minimal error. This calculation is performed continuously, generating a smooth sequence of joint space trajectories over the entire 6.5-second range. The joint angle update frequency is maintained at 50 Hz, and the output is the joint angle value and angular velocity at each moment.

[0050] Two types of data are fused into auxiliary output parameters. One type is the joint motion trajectory data obtained by S244, namely, the six joint angle values ​​and angular velocities at each timestamp. For example, at 4.00s, the shoulder joint angles are (0.80rad, 0.20rad, 0.35rad), the elbow extension is 0.45rad, and the wrist joint is (-0.15rad, 0.10rad). The other type is the dynamic contact force vectors obtained by S243, such as 9N for the left arm, 6N for the right arm, and 12N for the back at the same time. The data fusion method is to couple the joint trajectory sequence with the contact force vector group at each time point to generate an extended parameter matrix. For example, the matrix unit form is [joint angle, joint velocity, contact force magnitude, contact offset direction]. Taking 4.00s as an example, the left arm joint combined angle of 0.75rad corresponds to a contact force of 9N, and the coupled unit forms {0.75, 0.03rad / s, 9, -0.002m}. All joint data and contact forces are combined and output to form a parameter matrix of six degrees of freedom × contact points. The resulting auxiliary output parameter sequence encompasses a timeline from 0s to 6.5s, fully expressing both motion and force. This parameter package is directly transmitted to the nursing robot's execution controller, driving the robotic arm to perform precise assistive maneuvers that match the patient's real-time dynamic state during the lifting process.

[0051] Step S3 includes the following steps: Step S31: performing parameter feature importance learning on the robot auxiliary output parameters to obtain auxiliary learning parameters; Step S32: performing instruction encoding according to the auxiliary learning parameters to obtain auxiliary parameter instructions; Step S33: using the auxiliary parameter instructions as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture; Step S34: Send the robot-assisted nursing framework to the control terminal of the nursing robot to execute the control of the nursing robot.

[0052] In an embodiment of the present invention, the robot-assisted output parameter sequence generated in step S24 contains multiple dimensional data, such as joint angle, angular velocity, contact force vector, and contact direction correction value. In order to perform importance analysis, the parameters need to be expanded in the time dimension. For example, at 3.80s, the left arm contact force is 9N, the joint angle is 0.75rad, the joint angular velocity is 0.03rad / s, and the back contact force is 12N. By statistically analyzing the parameters within the entire 6.5s period, the contribution of each type of parameter to the overall trend is first calculated. The method is: taking the center of gravity offset trend as the target factor, and calculating the correlation values ​​of different parameters and targets respectively. It was found in the experiment that the correlation coefficient between the contact force vector and the center of gravity offset reached 0.87, while the correlation coefficient between the joint velocity and the center of gravity offset was only 0.54. Then, through the variance contribution verification, the variance of the longitudinal contact force is 2.3 (N²), accounting for 41% of the overall index. It is thus obtained that the importance weight of the contact force parameter is the largest, followed by the joint angle, and then the angular velocity. The final output auxiliary learning parameters are expressed as a set of weight coefficients, such as contact force = 0.45, joint angle = 0.35, angular velocity = 0.20, which serves as the input for subsequent steps.

[0053] After completing the feature importance analysis, the imperative encoding needs to be generated. The process is as follows: For each parameter category, encoding is performed according to the weight assigned in S31. For example, in joint angle and angular velocity control, weights of 0.35 and 0.20 are used to adjust them, combining the two input categories into an angle control factor. In contact force vector control, a weight of 0.45 is used to emphasize its priority in the support process. The encoding rule uses integerization, mapping each parameter to a fixed-width encoding bit. For example, contact force parameters in the range of 0–30 N are mapped to integer values ​​from 0–255, and joint angular velocity parameters in the range of 0–0.2 rad / s are mapped to integer values ​​from 0–255. For example, in a real experiment, at time 4.00 s, the left arm contact force is 9 N, which is encoded as 77; the right arm contact force is 6 N, which is encoded as 51; and the back contact force is 12 N, which is encoded as 102. A joint angular velocity of 0.03 rad / s is encoded as 38, and the corresponding joint angle of 0.75 rad is encoded as 136. The final combination forms a set of instruction data packets, such as [77,51,102,38,136]. This data packet is the auxiliary parameter instruction, representing the action posture and force distribution that the robot needs to perform at that moment.

[0054] Auxiliary parameter commands serve as input to the nursing robot's control architecture. First, the command elements are decomposed into three types of factors: functional factors, timing factors, and constraint factors. Functional factors are motion commands for different joints, such as shoulder lift, elbow extension, and wrist deflection. Timing factors specify the refresh rate, set at 50 Hz in this experiment. Constraint factors define the mechanical range, such as force output not exceeding 30 N and angular velocity not exceeding 0.2 rad / s. These factors are then combined to form a primary control matrix. For example, at time 4.00 s, the functional factor matrix is ​​{shoulder lift 0.75 rad, force output 9 N; elbow extension 0.45 rad, force output 6 N; back support position 0.012 m, force output 12 N}. This matrix is ​​logically superimposed to form a control framework, which is further integrated with the timing factors to produce an hourly updated auxiliary nursing architecture. Finally, priority rules are set within the architecture, such that longitudinal support commands take precedence over lateral correction commands. When there is a conflict between the longitudinal direction and the horizontal deflection, the constraint factors ensure that control resources are allocated to the longitudinal motion first. This design ensures stable execution of the architecture under complex robot movements.

[0055] In the final step, the generated care framework needs to be sent to the robot control terminal for implementation. The specific process involves first converting the care framework generated in S33 into a communication data packet, which includes joint control sub-commands and contact force control sub-commands. Each data packet contains a timestamp (with 1ms accuracy), target position parameters (in meters), target angle parameters (in rad), and target force parameters (in Newtons). For example, the data packet at time 4:00s contains: shoulder joint target angle 0.75 rad, force 9 N; elbow joint target angle 0.45 rad, force 6 N; back contact target displacement 0.012 m, force 12 N. The data packet is transmitted to the robot control terminal via the bus at a frequency of 50Hz. After receiving the data, the terminal directly drives the actuators, including the motor controller, force sensor feedback module, and end-user manipulator. The control terminal samples the feedback signal at a millisecond rate and compares it in real time with the target values ​​specified in the care framework input, making minor corrections as necessary. This entire process ensures that the robot arm performs continuous, stable, and precise assistive movements during the 6.5s it takes for the patient to rise.

[0056] Step S33 includes the following steps: Step S331: taking the auxiliary parameter instruction as input, performing instruction element decomposition processing on the auxiliary parameter instruction to obtain a multi-dimensional instruction factor set including instruction function factors, timing factors, and constraint factors; Step S332: deriving the functional interaction coupling relationship of the nursing robot based on the multi-dimensional instruction factor set to obtain the functional interaction coupling relationship; Step S333: performing policy association learning on the functional interaction coupling relationship according to the policy gradient algorithm to obtain an interaction coupling association strategy; Step S334: Design a nursing robot control architecture based on the interactive coupling association strategy to build a robot-assisted nursing architecture.

[0057] In this embodiment of the present invention, the input is the auxiliary parameter instruction obtained in step S32. This instruction is a combination of data encoded from the contact force vector, joint angle, angular velocity, and timestamp. For example, at time 4.00s, the instruction format is [9N, 6N, 12N, 0.75rad, 0.45rad, 0.03rad / s, 4000ms]. To construct a multidimensional factor set, the composite instruction needs to be decomposed into three types of elements: functional factors, timing factors, and constraint factors. Functional factors describe the specific execution task, such as lifting the shoulder joint, supporting the back, and lifting the forearm; timing factors specify the refresh rate, which in this experimental condition is 50Hz, corresponding to a period of 0.02s; and constraint factors specify physical limitations, such as single joint output force ≤ 30N, joint angular velocity ≤ 0.2rad / s, and maximum joint angle ≤ 2.0rad. The decomposition process maps the original encoded data into a three-dimensional vector to form a functional matrix (such as joint target angle and applied torque), a timing matrix (such as periodic sampling time interval), and a constraint matrix (such as safe torque limit). Taking the time 4.00 seconds as an example, the functional factors are defined as a shoulder angle of 0.75 rad and a force of 9 N, an elbow angle of 0.45 rad and a force of 6 N, and a back translation of 0.012 m and a force of 12 N. The timing factor is the refresh interval of 0.02 seconds, and the constraint factor is the output force limit of 30 N. This forms a complete set of multi-dimensional instruction factors.

[0058] The multidimensional factor set obtained in S331 serves as the basic input. During the nursing process, multiple functional factors influence each other. For example, there is a mechanical coupling between shoulder lift and back support, and an angular linkage between forearm lift and elbow rotation. The method for deriving functional interaction relationships is to first establish a dependency matrix between the factors. Taking a 9N shoulder lift and a 12N back support as an example, if the two forces act at a 10° angle, the coupling ratio of their mechanical components is cos(10°) = 0.985, indicating that the two forces are nearly oriented in the same direction and require synchronous coordination. Furthermore, at the temporal level, if the target shoulder angle update at time 3.95s is delayed by 0.01s, this will cause a misalignment between the force output and the back support action, resulting in a temporal coupling coefficient of 0.5 in the dependency matrix. Finally, an interaction matrix is ​​formed, in which each cell value represents the coupling strength between factors, ranging from 0 to 1. For example, the shoulder-back coupling strength is 0.85, the elbow-forearm coupling strength is 0.78, and the shoulder-elbow coupling strength is 0.62. The functional interaction coupling relationship obtained in this way can clearly express the interaction rules of different instruction factors under spatial mechanics and time constraints.

[0059] The functional interaction coupling relationship obtained in step S332 is further converted into an operational strategy in the control execution process. First, the input must be clarified, that is, the interaction coupling matrix, which is composed of the interaction strength between multidimensional factors. For example: in the measured process (standing up time period 3.50s~4.20s), the coupling coefficient between shoulder joint lifting and back support is 0.85, the coupling coefficient between elbow joint extension and forearm lifting is 0.78, and the coupling coefficient between shoulder joint and elbow joint is 0.62. Each coefficient represents the dependence strength of the corresponding functional factor in the execution of the action. In strategy association learning, it is necessary to construct an offset error function, and the difference between the actual displacement of the center of gravity and the target fitting displacement is used as an optimization indicator to correct the control signal. The error function is defined as: ,in , , Represent the expected center of gravity displacement values ​​in the x, y, and z directions respectively; , , represents the actual center of gravity displacement; E(t) is the error value of the center of gravity at time t, Indicates the time parameter. In the process of association learning, for each coupling pair (i, j), the output weight is defined Weights are incremented or decremented using a gradient to ensure the error function decreases. For example, when the center of gravity exhibits a lateral offset of 0.014m at 3.90s, the coupling matrix indicates a relationship strength of 0.85 between the shoulder joint and the back support. This means that shoulder motion compensation must be linked to the longitudinal force output of the back support. Through policy updates, the shoulder joint force is corrected from 9.0N to 9.5N, reducing the lateral offset to 0.005m. The policy learning process is performed frame by frame, with each update cycle lasting 0.02s, corresponding to a refresh rate of 50Hz. Experimental data shows that between 3.50s and 4.50s, the longitudinal center of gravity error decreases from 0.030m to 0.008m, and the lateral offset error decreases from 0.014m to 0.004m. Finally, after full data iteration, the output interactive coupling policy is a set of priority weights: longitudinal support control weight 0.45, lateral correction control weight 0.35, and joint velocity correction weight 0.20.

[0060] Using the interactive coupling strategy derived from S333, the control architecture was designed as a hierarchical control structure. The first layer, the center of gravity stabilization layer, primarily handles longitudinal support movements, ensuring that the body's center of mass moves upward synchronously with standing up. The second layer, the offset correction layer, allocates corrective movements to the elbow and shoulder joints in real time based on lateral and fore-aft offset trends. The third layer, the constraint safety layer, monitors all joint and force outputs to ensure they are within safety boundaries, immediately limiting them if they exceed preset limits. In the actual experiment, the auxiliary control commands at 4.00s were: the back actuator applies a 12N thrust, the shoulder joint applies a 9N lifting force, and the elbow joint maintains a 0.45rad angle. Due to the priority set by the coupling association strategy, the back thrust signal is immediately executed through the primary control channel, while the elbow joint signal undergoes offset correction in the secondary channel before being output. Timing control is continuously executed at a 50Hz refresh rate, ensuring that all joint outputs and auxiliary forces are updated within a 20ms period. Ultimately, the robot-assisted care architecture is constructed as a hierarchical processing structure, including functional factors, time series parameters, and constraint detection modules for each control channel, fully implementing the control mapping from strategy to execution.

[0061] The present invention also provides a control system for a nursing robot, which is used to execute the control method for the nursing robot as described above. The control system for the nursing robot includes: The data acquisition module is used to collect the posture trajectory of the patient when standing up during the nursing process through a 3D camera; collect muscle tension change signals of the body contact area when the patient stands up during the nursing process through an electromyographic signal sensor and based on the posture trajectory; and synchronously correlate the muscle tension change signals based on the posture trajectory to obtain posture-related muscle tension change data; The robot output parameter analysis module is used to analyze the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data. Subsequently, it performs center of gravity offset regression deconstruction to obtain the center of gravity offset regression trend; and generates robot-assisted output parameters based on the center of gravity offset regression trend. The control architecture design module is used to take the robot-assisted output parameters as input to design the control architecture of the nursing robot, so as to build a robot-assisted nursing architecture, and send it to the control terminal of the nursing robot to execute the control of the nursing robot.

[0062] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A control method for a nursing robot, characterized in that: An electromyographic signal sensor is deployed on the palm of the robotic arm of the nursing robot, and a 3D camera is deployed in the head area of ​​the nursing robot. The control method of the nursing robot includes the following steps: Step S1: using a 3D camera to collect a patient's posture trajectory when the patient stands up during the nursing process; using an electromyographic signal sensor and based on the posture trajectory, collecting a muscle tension change signal of the body contact area when the patient stands up during the nursing process; performing synchronous correlation processing on the muscle tension change signal based on the posture trajectory to obtain posture-correlated muscle tension change data; Step S2: analyzing the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data, and then performing center of gravity offset regression deconstruction to obtain a center of gravity offset regression trend; generating robot-assisted output parameters based on the center of gravity offset regression trend; Step S3: The robot-assisted output parameters are used as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture, and the architecture is sent to the control terminal of the nursing robot to execute the control of the nursing robot.

2. The control method of the nursing robot according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: using a 3D camera to capture the patient's posture trajectory when he or she stands up during the nursing process; Step S12: Using an electromyographic signal sensor and based on the posture trajectory, collecting muscle tension change signals of the patient's body contact areas when the patient stands up during the nursing process; the body contact areas include: the hand / forearm area, the back and shoulder blade area, and the elbow / upper arm area; Step S13: filtering the muscle tension change signal for noise, and then embedding the timestamp to obtain a change signal time series; Step S14: performing synchronous correlation processing on the time series of the change signal based on the posture trajectory to obtain posture-related muscle tension change data.

3. The control method of the nursing robot according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: constructing the spatial coordinates of the posture trajectory during the standing up process to obtain the spatial coordinates of the posture trajectory; Step S22: analyzing the muscle group force line imbalance strength during the posture change process on the posture-related muscle tension change data, thereby obtaining the force line imbalance strength; Step S23: performing gravity center offset regression deconstruction on the force line imbalance intensity based on the posture trajectory spatial coordinates to obtain a gravity center offset regression trend; Step S24: Generate robot-assisted output parameters based on the center of gravity offset regression trend, combined with the posture trajectory space coordinates and posture-related muscle tension change data.

4. The control method of the nursing robot according to claim 3, characterized in that: Step S22 includes the following steps: Step S221: calculating the average difference in tension intensity changes during posture changes based on the posture-associated muscle tension change data; Step S222: performing a three-dimensional force line direction vector decomposition of the posture-related muscle tension change data according to the tension intensity change mean difference during the posture change process to obtain a spatial force line distribution vector set; Step S223: performing orthogonal second-order statistical decomposition on the spatial force line distribution vector set to obtain force line second-order decomposition data; Step S224: analyzing the imbalance strength of muscle group force lines during posture changes using the second-order force line decomposition data on the spatial force line distribution vector set, thereby obtaining the force line imbalance strength.

5. The control method of the nursing robot according to claim 3, characterized in that: Step S23 includes the following steps: Step S231: Deconstructing the spatial coordinates of the posture trajectory to obtain the speed change rate, joint motion angles, and trunk posture angles when the patient stands up during the nursing process; the joints include the hip joint, knee joint, ankle joint, shoulder joint, and elbow joint; and the trunk refers to the area from the pelvis to the shoulder. Step S232: obtaining a dynamic center of gravity trajectory based on the velocity change rate, joint motion angle, and trunk posture angle; performing a center of gravity displacement vector polynomial fitting on the dynamic center of gravity trajectory to generate a center of gravity displacement fitting vector; Step S233: performing coordinate mapping processing on the force line imbalance intensity based on the posture trajectory spatial coordinates to generate local force line imbalance distribution data corresponding to each spatial coordinate point; Step S234: performing gravity center trend vector differentiation on the local force line imbalance distribution data based on the gravity center displacement fitting vector to obtain a gravity center offset trend sequence; Step S235: performing correlation regression analysis on the center of gravity shift trend sequence, and then performing center of gravity shift regression deconstruction to obtain the center of gravity shift regression trend.

6. The control method of the nursing robot according to claim 3, characterized in that: Step S24 includes the following steps: Step S241: Determine the patient's posture orientation when standing up based on the center of gravity offset regression trend and the posture trajectory spatial coordinates; Step S242: performing fatigue load analysis on the posture-related muscle tension change data to determine the acceptable muscle force margin; Step S243: Optimizing multi-point contact mechanics based on the posture orientation and the muscle acceptable force margin to obtain dynamic contact force; Step S244: performing adaptive fuzzy inverse kinematics transformation of the robot based on the dynamic contact force and the posture orientation to obtain the robot joint space motion trajectory; Step S245: performing data parameter fusion based on the robot joint spatial motion trajectory and dynamic contact force to generate robot auxiliary output parameters.

7. The control method of the nursing robot according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing parameter feature importance learning on the robot auxiliary output parameters to obtain auxiliary learning parameters; Step S32: performing instruction encoding according to the auxiliary learning parameters to obtain auxiliary parameter instructions; Step S33: using the auxiliary parameter instructions as input to design a nursing robot control architecture to construct a robot-assisted nursing architecture; Step S34: Send the robot-assisted nursing framework to the control terminal of the nursing robot to execute the control of the nursing robot.

8. The control method of the nursing robot according to claim 7, characterized in that: Step S33 includes the following steps: Step S331: taking the auxiliary parameter instruction as input, performing instruction element decomposition processing on the auxiliary parameter instruction to obtain a multi-dimensional instruction factor set including instruction function factors, timing factors, and constraint factors; Step S332: deriving the functional interaction coupling relationship of the nursing robot based on the multi-dimensional instruction factor set to obtain the functional interaction coupling relationship; Step S333: performing policy association learning on the functional interaction coupling relationship according to the policy gradient algorithm to obtain an interaction coupling association strategy; Step S334: Design a nursing robot control architecture based on the interactive coupling association strategy to build a robot-assisted nursing architecture.

9. A control system for a nursing robot, characterized in that: For executing the control method of the nursing robot according to claim 1, the control system of the nursing robot comprises: The data acquisition module is used to collect the posture trajectory of the patient when standing up during the nursing process through a 3D camera; collect muscle tension change signals of the body contact area when the patient stands up during the nursing process through an electromyographic signal sensor and based on the posture trajectory; and synchronously correlate the muscle tension change signals based on the posture trajectory to obtain posture-related muscle tension change data; The robot output parameter analysis module is used to analyze the imbalance strength of muscle group force lines during posture changes based on the posture-related muscle tension change data. Subsequently, it performs center of gravity offset regression deconstruction to obtain the center of gravity offset regression trend; and generates robot-assisted output parameters based on the center of gravity offset regression trend. The control architecture design module is used to take the robot-assisted output parameters as input to design the control architecture of the nursing robot, so as to build a robot-assisted nursing architecture, and send it to the control terminal of the nursing robot to execute the control of the nursing robot.

Images